Issue 3210: Inheritance of Stereotypes (uml-rtf)
Source: David Frankel Consulting (Mr. David Frankel, david(at)dfrankelconsulting.com)
Nature: Uncategorized Issue
Severity: 
Summary: Issue: The semantics of inheritance (UML 1.3,  ad/99-06-8, p. 60-61) does
not include Stereotypes as being among the things that a subclass inherits
from its superclasses. 

Recommendation: Explicitly specify that Stereotypes are inherited.

Discussion: UML specifies that Constraints are inherited.  A Stereotype is
logically a constraint, and can even formally define Constraints that apply
to stereotyped ModeledElements.  Consider a Class A that is stereotyped S, where C is the Constraint that the Stereotype definition specifies for all ModelElements stereotyped S.  Now consider Class B that
inherits A.  It stands to reason that the Constraint C applies to Class B.
Resolution: 
Revised Text: Stereotypes are GeneralizableElements and a discussion of what it means to subtype a Stereotype and what the applicable rules are has been added.

Actions taken:
January 11, 2000: received issue
May 24, 2001: closed issue
Discussion: 

End of Annotations:=====
Sender: dfrankel@gendev.com
Message-Id: <Version.32.20000111102210.00f70850@gendev.vwh.net>
X-Sender: dfrankel@gendev.vwh.net
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.0
Date: Tue, 11 Jan 2000 17:10:21 -0800
To: uml-rtf@omg.org
From: David Frankel <dfrankel@gendev.com>
Subject: Issue: Inheritance of Stereotypes
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
X-UIDL: $,Ke9/Ocd9;a'e9_&bd9

Issue: The semantics of inheritance (UML 1.3,  ad/99-06-8, p. 60-61)
does
not include Stereotypes as being among the things that a subclass
inherits
from its superclasses. 

Recommendation: Explicitly specify that Stereotypes are inherited.

Discussion: UML specifies that Constraints are inherited.  A
Stereotype is
logically a constraint, and can even formally define Constraints that
apply
to stereotyped ModeledElements.  Consider a Class A that is
stereotyped
<<S>>, where C is the Constraint that the Stereotype definition
specifies
for all ModelElements stereotyped <<S>>.  Now consider Class B that
inherits A.  It stands to reason that the Constraint C applies to
Class B.  



=============================
David S. Frankel
Chief Scientist
Genesis Development Corporation
741 Santiago Court
Chico, CA 95973-8781 U.S.A.

http://www.gendev.com
dfrankel@gendev.com
+1-530-893-1100 phone
+1-530-893-1153 fax
=============================

From: BELAUNDE Mariano CNET/DTL/LAN
	 <mariano.belaunde@cnet.francetelecom.fr>
To: uml-rtf@omg.org
Subject: RE: Issue: Inheritance of Stereotypes
Date: Wed, 12 Jan 2000 14:33:58 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2448.0)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org
id IAA01660
Content-Type: text/plain;
	charset="iso-8859-1"
X-UIDL: :DT!!RCV!!'M\d9F<5!!

I agree with David's issue.
Notice that it stands also to reason that inheritance applies to
pseudo-attributes 
(tagged values)
The recommendation associated with this issue can be completed in this
way:
Explicitly specify that Stereotypes are inherited as well as
any tagged value and constraint defined (required) in the stereotype.

Mariano Belaunde
France Telecom


-----Message d'origine-----
De: David Frankel [mailto:dfrankel@gendev.com]
Date: mercredi 12 janvier 2000 02:10
@: uml-rtf@omg.org
Objet: Issue: Inheritance of Stereotypes


Issue: The semantics of inheritance (UML 1.3,  ad/99-06-8, p. 60-61)
does
not include Stereotypes as being among the things that a subclass
inherits
from its superclasses. 

Recommendation: Explicitly specify that Stereotypes are inherited.

Discussion: UML specifies that Constraints are inherited.  A
Stereotype is
logically a constraint, and can even formally define Constraints that
apply
to stereotyped ModeledElements.  Consider a Class A that is
stereotyped
<<S>>, where C is the Constraint that the Stereotype definition
specifies
for all ModelElements stereotyped <<S>>.  Now consider Class B that
inherits A.  It stands to reason that the Constraint C applies to
Class B.  



=============================
David S. Frankel
Chief Scientist
Genesis Development Corporation
741 Santiago Court
Chico, CA 95973-8781 U.S.A.

http://www.gendev.com
dfrankel@gendev.com
+1-530-893-1100 phone
+1-530-893-1153 fax
=============================

From: MJCHONOLES@aol.com
Message-ID: <39.112bcce.25d3afa4@aol.com>
Date: Thu, 10 Feb 2000 01:07:32 EST
Subject: Comments on Issue 3210 -- Inheritance of Stereotypes...
To: uml-rtf@omg.org
CC: michael.j.chonoles@lmco.com
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Mailer: AOL 4.0 for Windows 95 sub 39
Content-Type: text/plain; charset="US-ASCII"
X-UIDL: [2Vd9~+!e9T@#!!OHE!!

While I'm not against inheriting stereotypes, I'm wondering about the 
complete extension of this to inheriting constraints and tagged-values
as 
indicated in the discussion. 

There needs to be exceptions for the constraints relating to
generalization. 
For example, the tagged value {abstract} should not be inherited. 

I believe the current UML already has some indications along this
line.

Other tagged values may also be better off not inherited, especially 
user-defined properties or constraints. While one could argue that
many 
common tagged values {author=Chonoles, dueDate=04/01/2000} should be 
inherited, some may be less appropriate, such as {obsolete},
{last-modified=12
/01/1999}.

We may need to consider, while updating the tailoring/extensibility 
capabilities of UML, the ability to specify whether a particular
extension 
(stereotype, property) is to be inherited.

Michael Jesse Chonoles
Chief of Methodology
Lockheed Martin Advanced Concepts Center
michael.j.chonoles@lmco.com

Date: Thu, 10 Feb 2000 13:31:34 -0500
From: Ed Seidewitz <seidewit@concept5.com>
Reply-To: seidewitz@concept5.com
Organization: Concept Five Technologies
X-Mailer: Mozilla 4.7 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: MJCHONOLES@aol.com
CC: uml-rtf@omg.org
Subject: Re: Comments on Issue 3210 -- Inheritance of Stereotypes...
References: <39.112bcce.25d3afa4@aol.com>
Content-Type: multipart/mixed;
 boundary="------------3A391DD6BF5F50C02EC450E8"
X-UIDL: ?V'e9"=d!!Y,"e9;k'e9

MJCHONOLES@aol.com wrote:

> While I'm not against inheriting stereotypes, I'm wondering about
the
> complete extension of this to inheriting constraints and
tagged-values as
> indicated in the discussion.
>
> There needs to be exceptions for the constraints relating to
generalization.
> For example, the tagged value {abstract} should not be inherited.

Note that the notation "{abstract}" actually refers to the attribute
"isAbstract"
of the class GeneralizableElement. As usual this attribute is
inherited by
descendant metaclasses of GeneralizableElement, but this does not mean
that the
VALUE of this metaattribute is "inherited" at the model level from a
modeled
generalizable element to its modeled descendants. This is why a
subclass of an
abstract class need not be abstract.

> I believe the current UML already has some indications along this
line.
>
> Other tagged values may also be better off not inherited, especially
> user-defined properties or constraints. While one could argue that
many
> common tagged values {author=Chonoles, dueDate=04/01/2000} should be
> inherited, some may be less appropriate, such as {obsolete},
{last-modified=12
> /01/1999}.

The use of the same model-level notation for tagged values and
metaattributes
indicates the intended role of tagged values as
"pseudo-metaattributes". This role
is clearest in the case that the tagged value is attached to a
stereotype. In this
case, it is reasonable to talk about the stereotype being "inherited",
meaning
that a child generalizable element should have the same stereotype (or
perhaps any
specialization of that stereotype) as its parent. With a stereotype so
"inherited", then the child element also seems to "inherit" the tagged
values of
that stereotype. However, just as in the case of metaattributes, the
VALUES of the
tagged values would not be "inherited", and so could be different in
the child
(unless there were some special semantic constraint associated with
the stereotype
that required such value conformance).

-- Ed

[] seidewit.vcf 

From: "Griss, Martin" <griss@exch.hpl.hp.com>
To: "'MJCHONOLES@aol.com'" <MJCHONOLES@aol.com>, uml-rtf@omg.org
Cc: michael.j.chonoles@lmco.com
Subject: RE: Comments on Issue 3210 -- Inheritance of Stereotypes...
Date: Thu, 10 Feb 2000 12:01:30 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
X-UIDL: @d;!!Vk<e9kCZ!!2,Q!!

We should also be able to specify non inheritance for attributes and
methods
for ordinary class/classifier inheritance, to ensure consistency of
the view
that
sterotyping is effectively addition of a (virtual) meta-class.


 
       Martin L. Griss,  Laboratory Scientist

PS. Excuse terseness, errors - I continue with my  RSI/tendonitis, and
have 
       started using a natural language system - less mistypes, more
mispokes.
							
       HP Laboratories,1U-16		ofc=650-857-8715
       1501 Page Mill Road,		fax=650-852-9841
       Palo Alto, CA 94301-1126    admin=650-236-2861 (Rhonda
Cummings)
       http://www.hpl.hp.com/reuse



-----Original Message-----
From: MJCHONOLES@aol.com [mailto:MJCHONOLES@aol.com]
Sent: Wednesday, February 09, 2000 10:08 PM
To: uml-rtf@omg.org
Cc: michael.j.chonoles@lmco.com
Subject: Comments on Issue 3210 -- Inheritance of Stereotypes...


While I'm not against inheriting stereotypes, I'm wondering about the 
complete extension of this to inheriting constraints and tagged-values
as 
indicated in the discussion. 

There needs to be exceptions for the constraints relating to
generalization.

For example, the tagged value {abstract} should not be inherited. 

I believe the current UML already has some indications along this
line.

Other tagged values may also be better off not inherited, especially 
user-defined properties or constraints. While one could argue that
many 
common tagged values {author=Chonoles, dueDate=04/01/2000} should be 
inherited, some may be less appropriate, such as {obsolete},
{last-modified=12
/01/1999}.

We may need to consider, while updating the tailoring/extensibility 
capabilities of UML, the ability to specify whether a particular
extension 
(stereotype, property) is to be inherited.

Michael Jesse Chonoles
Chief of Methodology
Lockheed Martin Advanced Concepts Center
michael.j.chonoles@lmco.com

From: MJCHONOLES@aol.com
Message-ID: <aa.1277b5e.25d4c14b@aol.com>
Date: Thu, 10 Feb 2000 20:35:07 EST
Subject: Re: Comments on Issue 3210 -- Inheritance of Stereotypes...
To: seidewitz@concept5.com
CC: uml-rtf@omg.org
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Mailer: AOL 4.0 for Windows 95 sub 39
Content-Type: text/plain; charset="US-ASCII"
X-UIDL: Q2gd9Il(!!9Fad9p/~e9

In a message dated 2/10/2000 1:30:52 PM Eastern Standard Time, 
seidewit@concept5.com writes:

<< The use of the same model-level notation for tagged values and 
metaattributes
 indicates the intended role of tagged values as
"pseudo-metaattributes". 
This role
 is clearest in the case that the tagged value is attached to a
stereotype. 
In this
 case, it is reasonable to talk about the stereotype being
"inherited", 
meaning
 that a child generalizable element should have the same stereotype
(or 
perhaps any
 specialization of that stereotype) as its parent. With a stereotype
so
 "inherited", then the child element also seems to "inherit" the
tagged 
values of
 that stereotype. However, just as in the case of metaattributes, the
VALUES 
of the
 tagged values would not be "inherited", and so could be different in
the 
child
 (unless there were some special semantic constraint associated with
the 
stereotype
 that required such value conformance).
 
 -- Ed
  >>
Yes, I see. Perhaps what needs to be clear is the need to be able
distinguish 
easily between inheritable constraints, meta-attributes, and
meta-attributes 
whose values are required to conform....

Michael Chonoles

From: Alan Moore <AlanM@artisansw.com>
To: "'MJCHONOLES@aol.com'" <MJCHONOLES@aol.com>,
seidewitz@concept5.com
Cc: uml-rtf@omg.org
Subject: RE: Comments on Issue 3210 -- Inheritance of Stereotypes...
Date: Fri, 11 Feb 2000 15:27:37 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
X-UIDL: )/\!!:"Ke9k_Wd9!R-e9

<<Do they both have to have the same
stereotype as a precondition in this case?>>

Actually, having thought about it, perhaps the specialized class
should
inherit the stereotype of the general class and add its own. How are
things
going with the notion of supporting multiple stereotypes anyway?

 -----Original Message-----
From:	Alan Moore [mailto:AlanM@artisansw.com] 
Sent:	11 February 2000 15:19
To:	'MJCHONOLES@aol.com'; seidewitz@concept5.com
Cc:	uml-rtf@omg.org
Subject:	RE: Comments on Issue 3210 -- Inheritance of
Stereotypes...

If one considers the case where a modeler has two classes (say) with
different stereotypes, what happens when the modeler tries to add a
Generalization relationship between them? Do they both have to have
the same
stereotype as a precondition in this case?

Also, in the case of multiple inheritance does the child have to have
multiple stereotypes if its parents have differing stereotypes?

Alan Moore
ARTiSAN Software Tools.

 -----Original Message-----
From:	MJCHONOLES@aol.com [mailto:MJCHONOLES@aol.com] 
Sent:	11 February 2000 01:35
To:	seidewitz@concept5.com
Cc:	uml-rtf@omg.org
Subject:	Re: Comments on Issue 3210 -- Inheritance of
Stereotypes...

In a message dated 2/10/2000 1:30:52 PM Eastern Standard Time, 
seidewit@concept5.com writes:

<< The use of the same model-level notation for tagged values and 
metaattributes
 indicates the intended role of tagged values as
"pseudo-metaattributes". 
This role
 is clearest in the case that the tagged value is attached to a
stereotype. 
In this
 case, it is reasonable to talk about the stereotype being
"inherited", 
meaning
 that a child generalizable element should have the same stereotype
(or 
perhaps any
 specialization of that stereotype) as its parent. With a stereotype
so
 "inherited", then the child element also seems to "inherit" the
tagged 
values of
 that stereotype. However, just as in the case of metaattributes, the
VALUES

of the
 tagged values would not be "inherited", and so could be different in
the 
child
 (unless there were some special semantic constraint associated with
the 
stereotype
 that required such value conformance).
 
 -- Ed
  >>
Yes, I see. Perhaps what needs to be clear is the need to be able
distinguish 
easily between inheritable constraints, meta-attributes, and
meta-attributes

whose values are required to conform....

Michael Chonoles

From: Alan Moore <AlanM@artisansw.com>
To: "'MJCHONOLES@aol.com'" <MJCHONOLES@aol.com>,
seidewitz@concept5.com
Cc: uml-rtf@omg.org
Subject: RE: Comments on Issue 3210 -- Inheritance of Stereotypes...
Date: Fri, 11 Feb 2000 15:19:04 -0000
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"
X-UIDL: Dl/!!L-C!!PH;!!>hWd9

If one considers the case where a modeler has two classes (say) with
different stereotypes, what happens when the modeler tries to add a
Generalization relationship between them? Do they both have to have
the same
stereotype as a precondition in this case?

Also, in the case of multiple inheritance does the child have to have
multiple stereotypes if its parents have differing stereotypes?

Alan Moore
ARTiSAN Software Tools.

 -----Original Message-----
From:	MJCHONOLES@aol.com [mailto:MJCHONOLES@aol.com] 
Sent:	11 February 2000 01:35
To:	seidewitz@concept5.com
Cc:	uml-rtf@omg.org
Subject:	Re: Comments on Issue 3210 -- Inheritance of
Stereotypes...

In a message dated 2/10/2000 1:30:52 PM Eastern Standard Time, 
seidewit@concept5.com writes:

<< The use of the same model-level notation for tagged values and 
metaattributes
 indicates the intended role of tagged values as
"pseudo-metaattributes". 
This role
 is clearest in the case that the tagged value is attached to a
stereotype. 
In this
 case, it is reasonable to talk about the stereotype being
"inherited", 
meaning
 that a child generalizable element should have the same stereotype
(or 
perhaps any
 specialization of that stereotype) as its parent. With a stereotype
so
 "inherited", then the child element also seems to "inherit" the
tagged 
values of
 that stereotype. However, just as in the case of metaattributes, the
VALUES

of the
 tagged values would not be "inherited", and so could be different in
the 
child
 (unless there were some special semantic constraint associated with
the 
stereotype
 that required such value conformance).
 
 -- Ed
  >>
Yes, I see. Perhaps what needs to be clear is the need to be able
distinguish 
easily between inheritable constraints, meta-attributes, and
meta-attributes

whose values are required to conform....

Michael Chonoles